Another core document required early in study start-up is the Investigator Site Manual. This document, distributed by the core lab to investigator sites, has a dual purpose: to establish exact imaging parameters and provide logistical assistance, such as shipping instructions to the site. The Investigator Site Manual typically contains far more detail on the requirements for submitting and collecting images for independent review than the clinical study protocol. Since it provides clear and concise requirements for imaging, the manual is a critical document for investigator sites.

Training and adequate monitoring are essential to assuring compliance with imaging guidelines and procedures set forth in the site manual. It is of no benefit to develop complex procedures without ensuring the investigator sites have the means and inclination to comply.

Data collection in an ideal world

Missing scans, unscheduled imaging visits, discrepant assessment dates, and discrepant exam types (e.g., spiral CTs versus conventional) create numerous problems with comparisons between the two analyses. It is often the case that differences of this sort are only discovered after an investigation into why RECIST assessments, for example, are discrepant. If discovered after an independent review, the data must be corrected and rereviewed—a costly and timely procedure.

Aside from radiological data, the clinical data utilized during the independent review process consistently becomes a source of complication during reviews. The sponsor's data management group is primarily responsible for receiving and redacting clinical information from the investigator sites for review by the independent radiologists and/or oncologists. In oncology trials where both radiologists and oncologists act as independent reviewers, clinical data can have a significant impact on the review results. Listing this data in the charter is critical for the core lab to know what type of data will be received. However, the type and organization of data often looks feasible during charter development, but very often becomes problematic at the time reviews are set to begin.

The problem with clinical data

Part of the problem with preparing and organizing clinical data for the independent review process is due to the way in which protocols are written. Study protocols are often developed without much forethought to collecting clinical information that easily facilitates independent review.

In oncology trials, data such as biopsy or cytology results often have a significant impact on how data are interpreted by the independent reviewers. An oncology dossier of predetermined clinical information from the sponsor's clinical CRF must be presented in a format that is relevant and unbiased.

Typically, the oncology review component of the review process involves an oncologist reviewing all radiological results in conjunction with a dossier for the subject's case, which covers the entire range of the subject's treatment during the trial. This type of review scenario is generally the preferred option in oncology trials. However, a dossier specified to be sectioned by timepoints/visits within a subject's case may be more scientifically sound than the global oncology dossier for certain disease indications.

In a chronic lymphocytic leukemia trial, where clinical data plays as large or a larger role than radiological data typically does in a solid tumor study, a dossier organized by timepoint may be more relevant to the clinician interpreting the data than a dossier for the subject's entire case. Similarly, in a melanoma trial, where skin lesion photographs can be reviewed and qualitatively assessed on a timepoint/visit basis, data may be more relevant if it is presented by timepoint. This timepoint by timepoint organization may sound reasonable in theory, but can present a challenge to the data management groups from the sponsor's side that will actually be organizing and exporting the data to the imaging core lab. The net result is that the charter must be amended, usually far too late in the process, because of limitations in the data collection process.